We are integrating the Genesys Cloud WebRTC softphone client (SDK v4.2.1) with a ServiceNow service portal via a custom AppFoundry widget. The goal is to trigger a screen pop in ServiceNow when an inbound voice interaction is answered, using the conversation:update webhook payload.
The issue occurs specifically when agents are logged into the UK West region. During the initial WebSocket handshake between the WebRTC client and the signaling server, we consistently receive a 408 Request Timeout error after approximately 3.5 seconds. The WebSocket connection never establishes, preventing the softphone from registering and subsequently blocking the downstream REST API call to ServiceNow.
We have verified the following:
- The integration user has the correct
web:readandweb:writescopes. - The CORS policy allows our ServiceNow instance domain.
- Network traces show the initial HTTP upgrade request reaches the Genesys edge but receives no response before the browser timeout.
Interestingly, this works flawlessly in the US East region. The UK West region appears to have stricter latency thresholds for the WebSocket upgrade phase. I suspect the 3.5s timeout is hardcoded in the SDK or the regional edge configuration, but I cannot find documentation on adjusting this threshold.
Has anyone encountered similar regional latency issues with WebRTC registration in UK West? Is there a configuration parameter in the connect method or the Architect flow that can extend the handshake timeout, or do we need to implement a retry mechanism in the AppFoundry code before attempting the ServiceNow API call?